The basic idea behind Cognitive Radio (CR) is to allow unlicensed users to utilize licensed frequency bands when they do not interfere to the primary (licensed) users. Thus, an important requirement of CR systems is to sense the spectrum in order to obtain awareness about the spectrum usage. In this paper, a new spectral estimation procedure for monitoring the radio spectrum which exploits frequency, time and angle diversity is presented. The procedure is a feature-based method able to detect predetermined spectral shape, providing at the same time an estimate of its power level, an estimate of its frequency location and an estimate of its angle of arrival. The specific spectral shape is called the candidate spectrum and gives name to the different methods.

Peer Reviewed