Sensing-Aided Precoding With High-Dynamic Moving Scatterers

Abstract

Sensing information can be leveraged to reduce the overheads associated with establishing and maintaining multiple-input and multiple-output (MIMO) communication links. Such information can be acquired from integrated sensing and communication (ISAC) capabilities. In this paper, we use sensing information to make precoding for spatial multiplexing more robust in high-dynamic environments with both quasi-static and mobile scattering objects. The significant variability of scattered paths requires frequent reconfigurations of the precoding matrices. To address this, we propose a sensing-aided precoding scheme for channels with high dynamics, leveraging the identification and localization of moving targets to concentrate energy mainly on slow-varying paths. Specifically, the geometric structure of the time-varying channel matrix is analyzed to uncover deviations in the location parameters of moving targets. A channel division criterion is then devised to partition the channel matrix into two matrices characterizing the paths of background scatterers and moving targets, respectively. Exploiting the digital-analog dual orthogonality between these matrices, precoding and combining matrices are jointly designed to suppress transmission leakage over mobile paths in both analog and digital domains. Numerical simulations validate that the proposed method achieves high spectral efficiency without requiring explicit channel tracking, thereby reducing complexity and overheads compared to conventional MIMO precoding methods.