Performance‐traffic‐trade‐off of two novel hard decision and two soft decision fusion periodogram‐based algorithms for cooperative spectrum sensing under unreliable reporting channel

Abstract

Spectrum sensing is one of the most important tasks of cognitive radios (CRs) to detect idle bands for opportunistic access. In centralised cooperative sensing, CRs sense a target band and share sensing information via reporting channel with a fusion centre to decide whether the sensed band is vacant or not. This information can be raw data in soft decision fusion scheme or local decisions on the channel occupation state in hard decision fusion one. A soft decision typically achieves better detection performances, but the hard one intrinsically offers smaller data traffic. This study proposes two periodogram-based hard decision algorithms and makes a detailed performance-traffic-trade-off with two periodogram-based soft-decision algorithms. All the approaches are based on the circular folding cooperative power spectral density split cancellation algorithm. Performance analyses are made under unavoidable errors in the reporting channel. In the first assumption, the simple memoryless binary symmetric channel is adopted. In the second one, the more realist memory Gilbert–Elliott channel is assumed. Results show that the proposed schemes are sensitive to reporting errors, but performance loss can be supplanted with channel encoding, yet keeping the traffic smaller than in the soft decision without channel encoding.