Cognitive ISAC Waveform Design Via Symbol-Level Precoding

This paper presents a symbol-level precoding (SLP) framework for cognitive integrated sensing and communication (ISAC) systems to address inter-cell interference in multi-user MIMO networks. Leveraging the cooperative nature of overlay cognitive radio (CR), the primary base station (PBS) shares user communication channel and data with the cognitive base station (CBS), enabling joint transmission and sensing over the same spectrum. The CBS simultaneously serves cognitive users (CUs) and performs radar sensing while ensuring coexistence with the PBS and its primary users (PUs). We design SLP-based cognitive waveforms that constructively exploit interference by minimizing radar beampattern error under communication QoS and per-antenna power constraints. This is a non-convex problem efficiently solved via manifold optimization using an augmented Lagrangian method. Simulation results show that the proposed cognitive SLP approach mitigates inter-cell interference and achieves notable gains in both sensing and communication compared to block-level precoding approach.