Low-complexity Leakage Minimization Beamforming for Large-scale Multi-user Cell-Free Massive MIMO

Abstract

We propose a low-complexity beamforming (BF) scheme for secrecy-rate maximization in multi-user (MU) cell-free massive multiple-input multiple-output (CF-mMIMO) systems, where legitimate users may act as non-colluding eavesdroppers of one another. To this end, we formulate an information leakage minimization problem and cast it into a tractable difference-of-convex algorithmic (DCA) form by leveraging fractional programming (FP). The resulting non-convex problem is solved through a concave-convex procedure (CCP)-based beamformer update, and an additional row-wise coordinate descent method (CDM) implementation is introduced to avoid explicit matrix inversion in the dominant linear-solve step. Additionally, we consider both direct transmit (TX)-BF and beyond-diagonal reconfigurable intelligent surface (BD-RIS)-assisted operation by defining an equivalent channel between each access point and user that combines the direct and reflective intelligent surface (RIS)-assisted propagation components. Simulation results show that the proposed secrecy-enhancement via leakage minimization (SecLM)-BF framework achieves secrecy and sum-rate performance close to state-of-the-art (SotA) semidefinite programming (SDP)-based benchmarks, as well as FP-based benchmarks, while providing a scalable inversion-free implementation for large-scale secure CF-mMIMO deployments.