Skip to content
Open access

Geometry-Adaptive Robust Beamforming for Movable-Antenna Arrays

2026 · IEEE Journal of Selected Areas in Sensors · Vol 3, pp. 554-567 · 0 citations · 37 references

Abstract

Movable-antenna arrays reshape their aperture by repositioning radiating elements, which adds spatial degrees of freedom that a fixed lattice cannot access. Worst-case robust beamforming is a natural tool for such arrays, but it relies on uncertainty margins fixed at design time. Those margins are no longer valid once the array response changes. This article ties the margins to the geometry by deriving them online from the noise-subspace structure of the current array covariance, so that they are recomputed whenever the elements move. Element positions and beamformer weights are then optimized jointly under an explicit reposition-cost penalty. The design is evaluated on a 32-element planar array supporting two users and two targets. Sensitivity to the initial margin choice falls by approximately 40 times relative to fixed margins. The advantage over fixed margins also grows with steering mismatch, from approximately 1.14 times at one degree to 2.00 times at five degrees, and the common-stream rate and the worst-target sensing SNR improve together. A priority weight that scales the two margin families traces a continuous rate-sensing tradeoff, whereas fixed margins admit a single operating point. Repositioning yields gains of 3.8% to 13.7% across five scenes under the evaluated conditions, and the largest gain occurs where the array has spatial slack rather than in the most degree-of-freedom-limited scene.

Read PDF

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.