COGNITIVE RADIO NETWORK FOR ROBUST SPECTRUM SENSING USING NEURAL NETWORKS

  • P Pavithra Roy et. al

Abstract

Cognitive Radio Networks (CRNs) for Spectrum sensing and allocation gained a lot of interest from the research community in recent years. Normally, spectrum sensing has been carried out using various prediction models and probabilistic algorithms available in the literature. Neural networks are also used to predict the spectrum availability, but it has drawbacks related to training algorithm employed. In this paper, a robust spectrum sensing algorithm using neural networks is proposed, which is a hybridized technique. Levenberg–Marquardt-based neural network is combined with a recent optimization algorithm, called Gravitational search algorithm so that sensing of the channel is improved. In this technique, the user data is transmitted through the idle or unoccupied channels that have been predicted using Gravitational search based Levenberg–Marquardt neural network (GS-LM) technique. The proposed technique is evaluated using SU, SUimp and throughput. A relative analysis was carried out by comparing proposed technique results to HMM, LM based NN and random technique.

Published
2020-02-02
How to Cite
et. al, P. P. R. (2020). COGNITIVE RADIO NETWORK FOR ROBUST SPECTRUM SENSING USING NEURAL NETWORKS. International Journal of Advanced Science and Technology, 29(04), 959 - 968. Retrieved from http://sersc.org/journals/index.php/IJAST/article/view/4765