Secure Multi-User Near-Field Transmission with Movable Antennas and a Stacked Intelligent Metasurface

This paper studies secure multi-user near-field transmission assisted by movable antennas (MAs) and a stacked intelligent metasurface (SIM) in the presence of multiple eavesdroppers. Each transmit antenna moves continuously along a common rail subject to aperture and minimum-spacing constraints, so its position independently controls one column of the near-field MA-SIM channel. We formulate the maximization of the secrecy sum rate against the strongest eavesdropper for each stream by jointly optimizing the transmit beamformers, MA positions, and SIM phase shifts. An alternating optimization algorithm combines a weighted minimum mean-square error (WMMSE) beamforming update, a layer-by-layer SIM phase update, and a successive convex approximation (SCA) update of all MA positions. Numerical results demonstrate that the proposed joint design outperforms the benchmarks, achieving secrecy sum rates more than 20% higher than those of the fixed uniform linear array (ULA) + SIM and MA-only schemes. The same-angle simulation further shows that spherical-wave range dependence enables positive secrecy rates when a plane-wave model yields zero, highlighting the role of near-field range discrimination.