A methodology for signals intelligence using non-conventional techniques and software-defined radio

In this paper, we present a methodology to conduct a performance evaluation of different spectral estimation techniques based on the probability of detection (Pd) by varying the signal-to-noise ratio (SNR). The spectral estimation methods are combined with an energy detector to detect radio signals transmitted in ultra-high frequency bands under higher noisy conditions. Traditionally, spectrum detection, a challenging task in signals intelligence, is performed in the frequency domain using the Fourier transform. However, other nonconventional techniques can be implemented, such as Burg, Yule-Walker, and Correlogram. As part of the methodology, a spectrum sensing system is implemented in GNU Radio, an open-source tool for software-defined radio applications. As a result of applying the proposed methodology, the spectrum sensing system based on the Correlogram can detect a simulated frequency modulated (FM) signal tuned to 462Mhz at even lower SNR. Under real FM signals, the system provided promising results.