Performance Analysis of Power Spectrum Entropy Detection for Unknown Sinusoidal Signals

Unknown signals detection under low signal noise ratio (SNR) condition was a problem of reconnaissance. The automatic detection algorithm based on power spectrum entropy was practical and effective，but the theoretical performance analysis had not been established yet. The receiver operating characteristic curve(ROC)of a power spectrum entropy detector was analyzed by statistical theory of detecting unknown sinusoidal signals in the background of white Gaussian noise. The theoretical results are basically consistent with the simulation. The performance comparison between power spectrum entropy detection and energy detection shows that the former performance is significantly better. The experimental results of detectable SNR in the sea are approximate to the theoretical prediction within 2 dB.