Distributed ADMM-Based Symbol-Level Precoding for Multi-Cell MIMO-OFDM ISAC

This letter studies symbol-level precoding (SLP) for multi-cell multi-input multi-output (MIMO) orthogonal frequency division multiplexing (OFDM) integrated sensing and communication (ISAC) under a coordinated beamforming (CBF) architecture. The goal is to maximize the target illumination power for sensing while meeting SLP constraints for communications. Unlike most multi-cell ISAC works in single-carrier settings, OFDM sensing couples the precoders both across base stations and across subcarriers, which complicates the design. To enable network-friendly computation, we develop a distributed alternating direction method of multipliers (ADMM) algorithm that uses only local channel state information and user data at each base station and exchanges a small set of consensus variables. We further propose a payload-reduced variant that aggregates inter-cell terms over subcarriers into a frequency-agnostic consensus, decoupling the exchanged payload from the number of subcarriers. A centralized successive upper-bound minimization method is provided as a benchmark. Simulations show that the distributed method converges quickly, approaches the centralized benchmark, and exhibits a predictable tradeoff in the payload-reduced setting.