A Covariance-Surrogate Framework for Movable-Antenna Enabled Anti-Jamming with Unknown Jammers

TL;DR

This paper introduces a surrogate-based optimization framework for movable-antenna anti-jamming systems with unknown jammers, improving SINR through joint antenna position and beamformer design.

Contribution

It proposes a novel surrogate objective and a low-complexity trust-region algorithm for joint antenna placement and beamforming in unknown jamming scenarios.

Findings

The surrogate objective closely approximates the true interference covariance.

The PTRSO algorithm converges to a stationary point near the anchor.

Numerical results show superior SINR performance over baselines.

Abstract

In this paper, we investigate a movable antenna (MA) enabled anti-jamming optimization problem, where a legitimate uplink system is exposed to multiple jammers with unknown jamming channels. To enhance the anti-jamming capability of the considered system, an MA array is deployed at the receiver, and the antenna positions and the minimum-variance distortionless-response (MVDR) receive beamformer are jointly optimized to maximize the output signal-to-interference-plus-noise ratio (SINR). The main challenge arises from the fact that the interference covariance matrix is unknown and nonlinearly dependent on the antenna positions. To overcome these issues, we propose a surrogate objective by replacing the unknown covariance with the sample covariance evaluated at the current antenna position anchor. Under a two-timescale framework, the surrogate objective is updated once per block (contains multiple snapshots) at the current anchor position, while the MVDR beamformer is adapted on a per-snapshot basis. We establish a local bound on the discrepancy between the surrogate and the true objective by leveraging matrix concentration inequalities, and further prove that a natural historical-averaging surrogate suffers from a non-vanishing geometric bias. Building on these insights, we develop a low-complexity projected trust-region (TR) surrogate optimization (PTRSO) algorithm that maintains the locality of each iteration via TR constraints and enforces feasibility through projection, which is guaranteed to converge to a stationary point near the anchor. Numerical results verify the effectiveness and robustness of the proposed PTRSO algorithm, which consistently achieves higher output SINR than existing baselines.