Cognitive radio network (CRN) came in to existence
as a promising solution to tackle issues due to scarcity of
spectrum. Spectrum sensing plays an important role for
maximizing the spectrum utilization where spectrum of the
primary users (PU) is sensed by the secondary user (SU) at
particular time and space. Researchers have presented machine
learning techniques for spectrum sensing, though, challenges
exists for the improvement in the throughput, energy efficiency,
detection probability and delivery ratio. In this paper, an
enhanced restricted Boltzmann machine (ERBM) is presented for
spectrum sensing based on RBM. Particle Swarm Optimization
(PSO) is incorporated for enhancing the performance of spectrum
sensing and computation of optimal momentum coefficient of
RBM. Simulation results shows that the performance of the
proposed spectrum sensing technique is comparable to the
existing techniques in terms of throughput, energy efficiency and
detection probability and delivery ratio.
Spectrum Prediction in Cognitive Radio Network Using Machine Learning Techniques
