Finite-Codebook RIS Beamforming for Direction-Separated ISAC: Phase-Error Analysis and Discrete Objective-Aware Design
Abstract
Finite-resolution phase control is a practical hardware constraint in reconfigurable intelligent surface (RIS)-aided integrated sensing and communication (ISAC), especially when communication and sensing directions are separated. This paper analyzes RIS phase quantization and develops a finite-codebook beamforming design. A phase-error coherence analysis characterizes how prescribed phase errors affect the coherent RIS-assisted term, per-element residual term, and direct–RIS interaction; under Rician fading, it explains why line-of-sight (LoS)-induced coherent gain is more sensitive to phase resolution than the diffuse component. Based on this structure, the proposed quantization-aware RIS alternating procedure (Q-RAP) combines a low-rank transmit beamformer update with quantization-aware discrete codeword search (Q-DCS), directly refining RIS phases within the finite-codebook feasible set without full-vector projection of continuous solutions. Numerical results validate the analysis, including biased and non-uniform phase-error cases, and show effective low-bit refinement, near-exhaustive small-size performance, small coordinate residuals across initializations, empirical robustness under imperfect channel state information (CSI), reduced projection-induced mismatch, and runtime scaling with array size and phase resolution.