Tensor Decomposition Based THz Channel Estimation in OTFS for Integrated Sensing and Communications

Due to the ultra-high transmission rate and resolution, integrated sensing and communications (ISAC) is promising to be achieved in terahertz (THz) frequency, which is treated as one of the most important technologies to enabled the development of autonomous systems in the coming sixth generation cellular communications (6G). In this paper, we investigate orthogonal time frequency space (OTFS) waveform-based ISAC in THz multi-input multi-output (MIMO) communications. To deal with the extremely high complexity in the traditional methods, we propose a tensor decomposition based THz channel estimation method in OTFS for ISAC with significantly low algorithm complexity. Then, the real-time requirement for the decision making based on ISAC in autonomous systems can be guaranteed. Specifically, we consider the THz downlink transmission from a base station (BS) to a mobile station (MS), where both of them are equipped with large-scale antenna arrays. OTFS modulation waveform is adopted for synchronous communication and sensing. Considering the sparse THz channel and the multidimension of the received multichannel signals, we propose an OTFS channel parameter estimation method based on CANDECOMP/PARAFAC (CP) decomposition, where the received signal is expressed as a third-order tensor. The tensor has a form of a low-rank CP decomposition, and satisfies the uniqueness of CP decomposition. Then, the estimated value of the channel parameter estimation can be obtained by the factor matrix obtained with CP decomposition. In addition, we analyze the algorithm complexity of the proposed method. Simulation results show the performance of the proposed method.