Time-Domain Zero-Padding (TZP) AFDM with Two-Stage Iterative MMSE Detection

In this paper, we investigate the design of low-complexity, high-performance transmission and detection schemes for the emerging affine frequency division multiplexing (AFDM) waveform in doubly selective channels. We first propose the time-domain zero-padding (TZP-)AFDM transmission frame, which simplifies time-domain input-output (IO) relation, while preserving the spectral efficiency and retaining a relatively compact discrete affine Fourier transform (DAFT)-domain IO relation via a phase-rotated overlap-add (PROLA) technique. Based on the derived IO relations, we develop a novel two-stage detector. In the first stage, low-complexity, initial linear minimum mean square error (MMSE) detection is performed in the time domain, with statistics efficiently computed and passed to the second stage. For the second stage, we introduce several iterative MMSE detection schemes that offer a trade-off between error performance and detection complexity. In particular, our cross-DAFT-and-time-domain iterative detector enhances error performance while exploiting the time-domain IO relation to reduce complexity. As an alternative, a DAFT-domain iterative detector is proposed for the second stage utilizing the DAFT-domain IO relation derived, which further improves the error performance albeit with increased complexity. Simulation results show that the proposed TZP-AFDM and detectors achieve superior error performance, higher spectral efficiency and lower computational overhead compared to existing approaches.