Spectrum sensing techniques for MANET simulations in discrete-time simulator NS-3

In recent years an emerging interest in cognitive networks has been identified, since it allows the use of licensed bandwidth by ordinary users. This functionality is embedded naturally in mobile ad-hoc networks (MANET), therefore research into cognitive MANET is attractive. There are several popular tools used for research in the field of MANET, one of them is the NS-3 discrete-time simulator. However, NS-3 currently does not provide support for cognitive network modelling. The key technology for cognitive networks to become feasible is the spectrum sensing detection of licensed user activity. To simulate cognitive MANETs using NS-3 it is necessary to implement spectrum sensing techniques in a software model. There are many algorithms, proposed to achieve robust spectrum sensing, differentiated by their performance and accuracy. The most robust are the algorithms based on the detection of cyclostationary features through the estimation of the cyclic spectra of the signal. This paper describes algorithms for cyclic spectra estimation, and proposes the method of integrating these algorithms into the NS-3 simulator.