Collaborative white space detection based on sample entropy and fractal theory

Distinguishing deterministic signal from noise in radio spectrum to detect white spaces for cognitive radio communication is vital task. To address this, quite a few sensing algorithms have been developed based on entropy measurement. However, most of them focused only on the information content in primary user transmitted signal and ignored the hidden complexity. Hence, in this work, the techniques that quantify hidden complexity in the signal rather than only information are studied using real-time Digital Television (DTV) signals. To quantify complexity, a test statistic is developed based on linear combination of sample entropy (SaEn(LC)) at different tolerance (r(t)) values. Furthermore, weighted collaborative detection method based on SaEn(LC) and fractal dimension measure is proposed to improve the detection accuracy by mitigating noise encountered by single user. The results reveal that the proposed method with five nodes can detect signals up to -23dB signal-to-noise ratio.