Cognitive radio frequency spectrum detection-The spectrum sensing of a wideband frequency range is studied by dividing it into multiple subbands.

It is assumed that in each subband either a primary user (PU) is active or absent in a additive white Gaussian

noise environment with an unknown variance. It is also assumed that at least a minimum given number of

subbands are vacant of PUs. In this multiple interrelated hypothesis testing problem, the noise variance is

estimated and a generalised likelihood ratio detector is proposed to identify possible spectrum holes at a

secondary user (SU). Provided that it is known that a specific PU can occupy a subset of subbands

simultaneously, a grouping algorithm which allows faster spectrum sensing is proposed. The collaboration of

multiple SUs can also be considered in order to enhance the detection performance. The collaborative

algorithms are compared in terms of the required exchange information among SUs in some collaboration

methods. The simulation results show that the propos