Scholarlink

<p>In Non-Orthogonal Multiple Access downlink networks, multiple signals are transmitted simultaneously for different users, often resulting in significant signal interference. Effective cancellation of these interfered-affected signals is essential for reliable communication. This research proposes a hybrid approach that integrates a Convolutional Neural Network with the African Buffalo Optimization algorithm to enhance the performance of Successive Interference Cancellation in downlink NOVA network. The CNN is used to model and predict interference patterns, while the ABO algorithm optimizes power allocation and synchronization parameters by leveraging bio-inspired search strategies. The proposed method is implemented in MATLAB and evaluated in terms of Bit Error Rate (BER), Mean Square Error (MSE), SIC loss, and sum rate. The proposed method achieves lower BER than CNN-SIC and conventional NOMA across the evaluated SNR range across an SNR range of 2.5–27.5 dB. The proposed method also improves sum rate by 3–15<b>%</b> compared with CNN-SIC (average <b>≈</b> 7.8%<b>)</b>, and by 3–24<b>%</b> compared with conventional NOMA (average <b>≈</b> 13.0%<b>)</b> across SNRs 2.5–27.5 dB. These results indicate that the proposed ABO-CNN improves signal reconstruction, interference cancellation performance, and network efficiency under the considered MATLAB simulation conditions.</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

Enhancement of Signal Interference Cancellation using Optimized Deep Learning ABO-CNN method in Downlink NOMA Network

  • 期刊论文

  • 发表日期：

    2026年7月20日

查看全文
  • Shyam Gehlot,
  • Harsh Pratap Singh,
  • Vedika Patil,
  • Nagendra Singh

摘要

In Non-Orthogonal Multiple Access downlink networks, multiple signals are transmitted simultaneously for different users, often resulting in significant signal interference. Effective cancellation of these interfered-affected signals is essential for reliable communication. This research proposes a hybrid approach that integrates a Convolutional Neural Network with the African Buffalo Optimization algorithm to enhance the performance of Successive Interference Cancellation in downlink NOVA network. The CNN is used to model and predict interference patterns, while the ABO algorithm optimizes power allocation and synchronization parameters by leveraging bio-inspired search strategies. The proposed method is implemented in MATLAB and evaluated in terms of Bit Error Rate (BER), Mean Square Error (MSE), SIC loss, and sum rate. The proposed method achieves lower BER than CNN-SIC and conventional NOMA across the evaluated SNR range across an SNR range of 2.5–27.5 dB. The proposed method also improves sum rate by 3–15% compared with CNN-SIC (average ≈ 7.8%), and by 3–24% compared with conventional NOMA (average ≈ 13.0%) across SNRs 2.5–27.5 dB. These results indicate that the proposed ABO-CNN improves signal reconstruction, interference cancellation performance, and network efficiency under the considered MATLAB simulation conditions.