Machine learning based spectrum decision in cognitive radio networks

The cognitive radio network (CRN) is considered as one of the promising solutions to address the issue of spectrum scarcity and eective spectrum utilization. In a CRN the Secondary User (SU) is allowed to occupy the spectrum which is temporarily not used by the Primary User (PU). Frequent interruptions from the PUs is the fundamental issue in CRN. The interruption forces SU to perform hando to another idle channel. On the other hand, spectrum hando can occur due to the mobility of the node. Hence, CRNs needs a smart spectrum decision scheme to timely switch the channels. An important issue in spectrum decision is spectrum hando. Since the SU’s spectrum usage is constrained by the PU’s trac pattern, it should carefully choose the right hando time. To increase the overall performance of the SU in the long term we use several machine learning algorithms in spectrum decision and compare it with the myopic decision which tries to achieve maximum performance in the short run.: