ABSTRACT
Spectrum sensing for acquiring the availability of the licensed spectrum band is com-
monly recognized as one of the most fundamental elements in cognitive radio-based
networks. Spectrum sensing can be realized as a two-layer mechanism [105]. On
the one hand, PHY-layer sensing focuses on efficiently detecting the signals of pri-
mary users to identify whether the primary users are present or not. Some PHY-layer
sensing methods including energy detection [43], matched filter [44], and feature de-
tection have been studied [45]. Moreover, some sequential spectrum sensing schemes
have also been investigated, such as the sequential shifted chi-square test [46], which
effectively reduces the average sample number compared with fixed-sample-size en-
ergy detection. On the other hand, MAC-layer sensing, which plays an important
role in a cognitive radio-based centralized network, determines the channels that sec-
ondary users should sense and access in each slot for good performance in terms of
sensing delays, throughput of secondary users, and so forth. Sensing delays result
in performance degradation of secondary users, especially in broadband communi-
cations systems since more sensing time is needed to acquire enough spectrum op-
portunities. Therefore, an important issue in MAC-layer sensing is how to acquire
more spectrum opportunities quickly. Collaborative spectrum sensing has also been
well studied in cognitive radio-based centralized networks in recent years since the
degradation of sensing performance caused by multi-path fading and shadowing can
be effectively overcome via collaborative sensing. Moreover, the sensing time can
also be reduced via collaborative sensing since more sensing data can be obtained
simultaneously.
