Abstract
Spectrum prediction based sensing schemes minimize the overall energy consumption of the sensing module in cognitive radio networks (CRNs) by predicting the status of spectrum before performing actual physical sensing. But, the performance of independent or local prediction models suffer from inaccuracies. Cooperative mode of spectrum prediction is found to be suitable to overcome the issues of local prediction models. In this work, we propose a cooperative spectrum prediction-driven sensing scheme for energy constrained cognitive radio networks to reduce the energy consumption while maintaining the spectral efficiency. The proposed scheme first employs a long short term memory network technique to perform local spectrum prediction, which identifies the status of a channel before actual sensing to improve energy efficiency. Thereafter, a parallel fusion based cooperative spectrum prediction model is applied to minimize the errors induced in local prediction model. Finally, the resultant cooperative prediction model is combined with a spectrum sensing framework to perform sensing operation when the cooperative spectrum prediction results to an indeterminate state in order to enhance the spectral efficiency. Simulation results show the efficacy of the proposed scheme in terms of spectral efficiency and energy efficiency compared to similar schemes from literature.
Highlights
With the advancement of technology, the usage of wireless devices increases enormously from the last decades [1]
The recent measurements carried out by Federal Communications Commission (FCC) have shown that 70% of the allocated spectrum in US is not utilized [2]. This motivates the development of the concept of cognitive radio (CR) [3], which allows CR enabled users or secondary users (SUs) to utilize the licensed radio spectrum when the spectrum is temporally not being utilized by its licensed users or primary users (PUs)
We applied an long short term memory (LSTM) based local spectrum prediction model to identify the status of a channel before actual sensing to improve energy efficiency
Summary
With the advancement of technology, the usage of wireless devices increases enormously from the last decades [1]. Due to the deteriorated sensing performance, more errors are introduced in the sensing dataset and as a result βe increases To capture such dynamics in network environment, we introduce an adaptive κ-out-of-N fusion rule (AF), which dynamically determines κ, that is, the number of SUs that take a cooperative decision at a particular time slot for a channel. In ad-hoc secondary network setup, the SUs can themselves play the role of FCs and distributed cooperative spectrum sensing [32] will best suit for this purpose but at the expense of message overhead, time, and extra energy at the SUs. While fusing the local prediction results, the cooperative model outputs two possible decision states, viz. SUs get to the sensing mode and perform spectrum sensing to determine the PU’s occupancy on the channel at a particular slot, when cooperative prediction outputs an IM state. Using AF-rule, the collaborative probabilities of false prediction (Qfp) and miss prediction (Qmp) are derived using (16) and (17) [19], respectively
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have