Low-Complexity Reliability-Based Equalization and Detection for OTFS-NOMA

TL;DR

This paper introduces a low-complexity, reliability zone-based detection method for OTFS-NOMA that outperforms traditional MMSE-SIC in high-mobility scenarios by effectively mitigating multi-user interference.

Contribution

It proposes a novel RZ detection scheme with optimized thresholds and an MSE-tracking approach to enhance OTFS-NOMA detection performance.

Findings

Superiority over MMSE-SIC in symbol error rate

Effective multi-user interference cancellation

Low-complexity iterative detection method

Abstract

Orthogonal time frequency space (OTFS) modulation has recently emerged as a potential 6G candidate waveform which provides improved performance in high-mobility scenarios. In this paper we investigate the combination of OTFS with non-orthogonal multiple access (NOMA). Existing equalization and detection methods for OTFS-NOMA, such as minimum-mean-squared error with successive interference cancellation (MMSE-SIC), suffer from poor performance. Additionally, existing iterative methods for single-user OTFS based on low-complexity iterative least-squares solvers are not directly applicable to the NOMA scenario due to the presence of multi-user interference (MUI). Motivated by this, in this paper we propose a low-complexity method for equalization and detection for OTFS-NOMA. The proposed method uses a novel reliability zone (RZ) detection scheme which estimates the reliable symbols of the users and then uses interference cancellation to remove MUI. The thresholds for the RZ detector are optimized in a greedy manner to further improve detection performance. In order to optimize these thresholds, we modify the least squares with QR-factorization (LSQR) algorithm used for channel equalization to compute the the post-equalization mean-squared error (MSE), and track the evolution of this MSE throughout the iterative detection process. Numerical results demonstrate the superiority of the proposed equalization and detection technique to the existing MMSE-SIC benchmark in terms of symbol error rate (SER).