Zak-OTFS ISAC With Bistatic Sensing via Semi-Blind Atomic Norm Denoising Scheme

Zak-transform-based orthogonal time frequency space (Zak-OTFS) modulation provides a delay-Doppler (DD) domain framework for integrated sensing and communication (ISAC) in high-mobility scenarios. Since the DD-domain channel response is directly related to the target parameters and also determines the communication link, accurate channel estimation is a key task for ISAC. However, it is challenging due to the fractional delay and Doppler shifts, which spread the channel response beyond the on-grid DD bins and lead to strong coupling between channel estimation and data detection. To address this issue, this paper proposes a semi-blind atomic norm denoising scheme for Zak-OTFS ISAC with bistatic sensing. We first derive the discrete-time input-output (I/O) relationship of Zak-OTFS with rectangular windowing. Based on this I/O relation, the joint channel parameter estimation and data detection problem is formulated as an atomic norm denoising problem, where a negative square penalty is introduced to handle the non-convex discrete constellation constraints. An accelerated iterative algorithm is then developed by combining majorization-minimization, accelerated projected gradient, and inexact accelerated proximal gradient methods. We also establish the convergence of the proposed algorithm. Simulation results show that the proposed scheme achieves super-resolution sensing accuracy and communication performance close to the perfect-CSI lower bound.