Skip to Main Content
Spectrum sensing is an essential part of future cognitive radios, as the spectrum sensor provides information about the utilization of the surrounding radio spectrum. Several approaches to implementing spectrum sensing have been proposed in the literature, one of them being the class of feature detectors. Cyclostationary feature detectors are, in general, considered to be superior in performance but suffer from high implementation complexity. Therefore, most studies still utilize energy detectors, which may not reach the performance requirements set for practical implementations. This paper presents angular domain feature detection algorithms that are based on cyclostationary properties. Angular domain signal processing is shown to simplify the implementation considerably while preserving comparable performance. Moreover, a new detection algorithm that leads to multiplier-free implementation and reduces the memory requirements, compared with any previous approaches, is proposed.