Abstract
Emerging technologies such as dynamic time division duplexing (D-TDD) in 5G and full-duplex (FD), both considered for 6G, promise higher spectral efficiency and lower latency. D-TDD flexibly allocates uplink and downlink resources to match traffic, while FD enables simultaneous transmission and reception on the same band, boosting capacity. These gains, however, introduce new interference: cross-link interference (CLI) in D-TDD and self-interference (SI) in FD, both of which threaten network robustness. Extra-large multiple-input multiple-output (XL-MIMO) with its near-field (NF) visibility-region (VR) concept offers a promising solution. The core idea in this work is to exploit VRs jointly in space and temporal domain to build a comprehensive interference-management framework. By modeling VR characteristics across spatial and temporal domains, we develop strategies that selectively activate antennas and suppress dominant interferers. We propose a VR-aware antenna-selection method that evaluates overlap and delay structure to minimize CLI and SI without sacrificing desired signal power. Simulation results validate the approach, showing substantial interference reduction and improved reliability and efficiency in XL-MIMO-based cellular systems.