Scholarlink

<p>This paper proposes a robust deep learning-assisted channel estimation framework for underwater acoustic (UWA) orthogonal frequency division multiplexing (OFDM) systems. The transmitter employs Henkel Function First Kind-employed Quadrature Phase Shift Keying (HFFK-QPSK) modulation and pulse-shaped filtering using a perfectly positive autocorrelated raised cosine filter (PPA-RCF) for spectral shaping and reduced inter-symbol interference. To address the time-varying Doppler distortions inherent in UWA channels, we implement a Cosine Probability Distributed Golden Search Optimized Covariance State Induced Kalman Filter (CP-GS-KF) algorithm for joint synchronization and Doppler compensation. For channel estimation, a novel Sparse-Aware Time-Reversal Artificial Neural Network (SATR-ANN) is designed and trained on synthetic sparse UWA channel data to learn nonlinear mappings between pilot symbols and channel responses. Furthermore, extensive simulation results demonstrate improved performance in terms of Bit Error Rate (BER), and Mean Square Error (MSE), validating the proposed system’s effectiveness for reliable and adaptive underwater communication in dynamic and challenging environments.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

An efficient underwater OFDM framework for Doppler-resilient and sparse channel estimation

  • 期刊论文

  • 发表日期：

    2026年5月18日

查看全文
  • Anand Kumar

摘要

This paper proposes a robust deep learning-assisted channel estimation framework for underwater acoustic (UWA) orthogonal frequency division multiplexing (OFDM) systems. The transmitter employs Henkel Function First Kind-employed Quadrature Phase Shift Keying (HFFK-QPSK) modulation and pulse-shaped filtering using a perfectly positive autocorrelated raised cosine filter (PPA-RCF) for spectral shaping and reduced inter-symbol interference. To address the time-varying Doppler distortions inherent in UWA channels, we implement a Cosine Probability Distributed Golden Search Optimized Covariance State Induced Kalman Filter (CP-GS-KF) algorithm for joint synchronization and Doppler compensation. For channel estimation, a novel Sparse-Aware Time-Reversal Artificial Neural Network (SATR-ANN) is designed and trained on synthetic sparse UWA channel data to learn nonlinear mappings between pilot symbols and channel responses. Furthermore, extensive simulation results demonstrate improved performance in terms of Bit Error Rate (BER), and Mean Square Error (MSE), validating the proposed system’s effectiveness for reliable and adaptive underwater communication in dynamic and challenging environments.