Near-Field Velocity Estimation and Predictive Beamforming with Modular Linear Array

Abstract: Velocity estimation is a cornerstone of recently introduced near-field predictive beamforming. This paper derives the closed-form Cramer-Rao bounds (CRBs) for joint velocity estimation using a modular linear array (MLA) within a predictive-beamforming framework. The analysis shows that increasing inter-module separation enlarges the effective aperture and reduces the transverse-velocity CRB, whereas the radial-velocity CRB is largely insensitive to separation. We further obtain a simple closed-form relation linking the achievable antenna savings to the inter-module separation while preserving the same transverse accuracy of a uniform linear array (ULA). We further investigate how velocity mismatch affects array gain and show that transverse-velocity errors cause more severe performance degradation than radial-velocity errors. Simulations show that predictive beamforming with MLAs maintains high localization accuracy for target tracking.