Estimation of spectrum sensing efficiency for optical OFDM for QAM schemes using SVM methods for VLC applications

This paper presents a comprehensive evaluation of spectrum sensing techniques for optical OFDM signals using various modulation formats, highlighting the superiority of the proposed support vector machine (SVM) algorithm. Performance is assessed in terms of probability of detection (Pd) versus signal-to-noise ratio (SNR) for 64-QAM, 256-QAM, and 512-QAM signals. For 64-QAM, the proposed SVM achieves a Pd of one at an SNR of 2 dB, significantly outperforming other methods such as conventional SVM, compressed sensing, matched filter, energy detection, and optical OFDM, which require higher SNRs ranging from 4 dB to 9.5 dB. Similar trends are observed for 256-QAM, where the proposed SVM achieves Pd = 1 at 6.5 dB, offering up to 8.5 dB SNR gain over other techniques. For 512-QAM, the proposed SVM achieves full detection at 9 dB, compared to 11 dB–16.5 dB for the others. These results demonstrate consistent SNR gains, confirming the robustness and reliability of the proposed method under low SNR conditions. Additionally, power spectral density (PSD) analysis reveals a significant reduction in spectral leakage, with the proposed method showing up to 500 W/MHz improvement in PSD. Overall, the proposed SVM approach is highly effective for spectrum sensing in high-order optical OFDM systems, offering both enhanced detection and spectral efficiency.