Scholarlink

<p> Cognitive Radio Networks (CRNs), spectrum sensing and allocation are essential due to the limited frequency bands and growing demand for wireless communication. However, the conventional spectrum sensing process is challenged since spectrum availability is dynamic and unpredictable. When the number of users and environmental conditions are changed, they may need extensive retraining. For effective spectrum sensing and allocation, this research proposes a Score test-based one-bit signal detector (ST1bSD) with an incremental ensemble support vector machine (SVM) model. First, in the presence of non-Gaussian noise, the principal user signal is identified using ST1bSD. Because of its straightforward structure, ST1bSD reduces the computational load and does not require prior knowledge about the transmitted signal. After that, a number of signal attributes are extracted as features, such as signal energy, power spectral density, signal-to-noise ratio (SNR), received signal strength indicator (RSSI), and goodness of fit (GoF). A Smart Aggregation Center combines these elements to produce a potent feature set for spectrum allocation. The innovative Machine Learning (ML) technique Incremental Ensemble SVM, which produces precise decisions for a constantly changing environment, is then given these features. The proposed method is simulated and compared with existing methods. An accuracy of 98.2% is achieved with the computation time of 0.0083&#xa0;s. The accuracy performance of the proposed approach is increased by 1.13% over the best existing method, namely Twin SVM. The experimental findings demonstrate that the proposed strategy performs better than other conventional strategies.</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

IncEML-CRNet: incremental ensemble learning for joint spectrum sensing and allocation in cognitive radio networks

  • 期刊论文

  • 发表日期：

    2026年2月12日

查看全文
  • S. Sadhana,
  • M. Vanitha

摘要

Cognitive Radio Networks (CRNs), spectrum sensing and allocation are essential due to the limited frequency bands and growing demand for wireless communication. However, the conventional spectrum sensing process is challenged since spectrum availability is dynamic and unpredictable. When the number of users and environmental conditions are changed, they may need extensive retraining. For effective spectrum sensing and allocation, this research proposes a Score test-based one-bit signal detector (ST1bSD) with an incremental ensemble support vector machine (SVM) model. First, in the presence of non-Gaussian noise, the principal user signal is identified using ST1bSD. Because of its straightforward structure, ST1bSD reduces the computational load and does not require prior knowledge about the transmitted signal. After that, a number of signal attributes are extracted as features, such as signal energy, power spectral density, signal-to-noise ratio (SNR), received signal strength indicator (RSSI), and goodness of fit (GoF). A Smart Aggregation Center combines these elements to produce a potent feature set for spectrum allocation. The innovative Machine Learning (ML) technique Incremental Ensemble SVM, which produces precise decisions for a constantly changing environment, is then given these features. The proposed method is simulated and compared with existing methods. An accuracy of 98.2% is achieved with the computation time of 0.0083 s. The accuracy performance of the proposed approach is increased by 1.13% over the best existing method, namely Twin SVM. The experimental findings demonstrate that the proposed strategy performs better than other conventional strategies.