Efficient 1-bit Block-Level Precoding for Massive MIMO–DFRC Systems

Abstract

This letter investigates the 1-bit block-level precoding (BLP) design for massive multiple-input multiple-output (MIMO) dual-function radar-communication (DFRC) systems under phase-shift keying (PSK) modulation. A discrete nonconvex problem is formulated by jointly maximizing the communication safety margin and matching a radar reference waveform. To solve this problem, an alternating optimization framework is proposed. The radar reference waveform is updated in closed form via a singular value decomposition (SVD)-based Procrustes projection. For a fixed radar reference waveform, the transmit signal matrix subproblem decomposes into independent symbol-level precoding (SLP) subproblems, for which a penalty-based bit-flip coordinate descent (bfCD) algorithm is developed. Simulations demonstrate that the proposed algorithm achieves competitive bit error rate (BER) performance and a flexible communication-radar tradeoff.