Blind spectrum sensing method for OFDM signal detection in Cognitive Radio communications

Abstract

Cognitive Radio is an enabling technology for accessing the unused spectrum. It may need to work in blind scenarios where it is unaware of the received signal parameters. In real-time military applications, the cyclostationary analysis of OFDM signals involves high computational complexity and requires additional processing and detection time. In order to detect the active carrier frequencies in such a scenario, we propose a blind two stage spectrum sensing scheme where the sequential sliding window energy detection is followed by cyclostationary feature detection that extracts the underlying periodic properties of the OFDM signal. The second-order cyclostationarity due to the equally spaced pilot subcarriers and due to the preamble with cyclic extension is explored. The peaks due to pilots and due to the preamble and cyclic extension are captured. The cyclostationary feature detection is performed over a selected cyclic spectrum instead of exploring the entire spectrum. The blind energy/cyclostationary detection of OFDM signals is compared with the matched filter based spectrum sensing algorithm of detecting OFDM signals. Simulations demonstrate the reliable and highly robust performance of the proposed non-parametric spectrum sensing method in Gaussian environment.