A Recursive Algorithm for Wideband Temporal Spectrum Sensing

Wideband spectrum sensing techniques determine which portions of a given spectrum band are occupied or idle in the frequency domain. The idle portions represent spectrum holes that can potentially be exploited by secondary or unlicensed users. Existing methods for wideband sensing, however, do not take into account the temporal activity of the primary or licensed users within the spectrum band. We propose an algorithm that identifies primary user activity over a wide spectrum band and provides a statistical characterization of the primary user signals in the band. The algorithm applies hidden Markov modeling to a hierarchically partitioned representation of the spectrum band, together with a recursive tree search. Different from existing wideband sensing algorithms, the proposed wideband temporal sensing method is able to accurately detect spectrum holes even in the presence of bursting primary user signals. Moreover, the hidden Markov modeling of the primary user signals enables the accurate detection and the prediction of primary user activity over time. Numerical results demonstrate the significant performance gain of the proposed algorithm over existing wideband spectrum sensing algorithms, particularly in the presence of low duty-cycle primary user signals.