Deep Learning for Signal Detection in Non-orthogonal Multiple Access Orthogonal Time-Frequency Space

Abstract

The Non-orthogonal Multiple Access Orthogonal Time-Frequency Space (NOMA-OTFS) modulation technique can address the multiuser communication challenges in high-mobility scenarios. To tackle the high model complexity in the downlink detection system for NOMA-OTFS modulation, this paper introduces a signal detection algorithm based on deep neural network technology. The use of deep learning techniques enhances the receiver model, reduces model complexity, and improves system efficiency. The algorithm’s stability, reliability, and scenario applicability are ensured through offline training of symbol-level datasets. Simulation results indicate that the proposed algorithm outperforms existing schemes in terms of Bit Error Rate (BER), demonstrating superior detection performance and robustness.