Year: 2026
Type: preprint
Abstract: Near-field arrays estimate position from wavefront curvature and relative received amplitudes. Unknown receiver phases and delays can obscure the phase information, making calibration a positioning design problem. We study which noisy calibration measurements are most useful when sensing and calibration share a finite received signal-to-noise ratio budget. Under a spherical-wave model with known relative receiver gains and known reference-path responses, we separate position information into amplitude and calibrated phase contributions. This yields a geometry condition for nonsingular local information and shows how reference frequencies constrain receiver phase at the sensing carrier. We then allocate resources between sensing and selected sensor pairs by convex optimization. Comparisons with uniform calibration designs and simulations of complex observations demonstrate the resulting accuracy gains. The framework distinguishes calibration needed for local identifiability from calibration needed for useful positioning accuracy.
Source: Research Square
Authors: Sajad Daei, Gabor Fodor, Mikael Skoglund
Institution: KTH Royal Institute of Technology
Language: English
Subfield: Signal Processing
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7234115186
