Quantizer Design for ISAC Systems Based on Mutual Information and Cramér–Rao Bound

To support 6G high-precision sensing and high-data-rate communications, integrated sensing and communication (ISAC) moves toward higher carrier frequencies and wider bandwidths, which places heavy demands on receiver ADCs and motivates low-resolution ADC designs. Nevertheless, most existing low-resolution ISAC works focus on 1-bit or multi-bit uniform quantization, with little attention paid to non-uniform threshold optimization for joint communication and sensing performance. This paper investigates multi-bit non-uniform quantizer design for OFDM-based ISAC receivers with low-resolution ADCs. The sensing performance is characterized by the Cramér–Rao bound (CRB) derived from quantized measurements, while the communication performance is measured by the mutual information between the transmitted symbols and the quantized observations. Based on the normalized communication and sensing metrics, a joint quantization-threshold optimization problem is formulated to balance sensing and communication performance. The resulting non-convex problem is solved using a successive convex approximation (SCA)-based iterative algorithm. Numerical results show that the proposed non-uniform quantizer approaches the unquantized benchmark with 5-bit ADC resolution. At the same 5-bit resolution, it increases the communication mutual information by approximately 21.58% and reduces the RCRB by approximately 47.66% compared with conventional uniform quantization. The proposed framework provides a multi-bit quantizer design approach for low-resolution ISAC receivers.