稳健性(进化)
计算机科学
收发机
电子工程
正交频分复用
无线
偏移量(计算机科学)
载波频率偏移
调制(音乐)
多路复用
限制
干扰(通信)
频率调制
绩效改进
人工智能
时频分析
频率偏移
转换器
噪音(视频)
UTC偏移量
频道(广播)
相位噪声
噪声测量
算法
相位调制
作者
Tonmoy Rajkhowa,Sanjeev Sharma,Kuntal Deka
标识
DOI:10.1109/lwc.2025.3632246
摘要
This letter presents the first comprehensive study on automatic modulation classification (AMC) for orthogonal time-frequency space (OTFS) systems under practical hardware-impairments (HIs), such as in-phase/quadrature imbalance (IQI), DC offset (DCO), phase noise (PN), and carrier frequency offset (CFO). Unlike prior OTFS-AMC studies that assume ideal transceivers, this work incorporates hardware non-idealities and systematically analyzes the individual and joint effects of multiple HIs on AMC performance under Extended Vehicular A (EVA) channel conditions. A 2D convolutional neural network (2D-CNN) model, tailored to capture the sparse structure of 2D OTFS frames, serves as a consistent baseline to assess OTFS robustness under practical HIs. Experimental results demonstrate that receiver-side HIs are more detrimental than transmitter-side. Moreover, it is observed that severe PN leads to severe AMC performance degradation, limiting the overall AMC accuracy to 11%. Furthermore, analysis also confirms that OTFS consistently outperforms conventional orthogonal frequency-division multiplexing (OFDM), which struggles with severe inter-carrier interference in doubly-spread wireless channels. Increasing the OTFS frame size enhances the robustness of AMC performance, providing a 3–5% accuracy improvement under various HI conditions. Collectively, these findings offer valuable design insights for developing HI-resilient OTFS-AMC systems suitable for practical high-mobility wireless networks.
科研通智能强力驱动
Strongly Powered by AbleSci AI