Hybrid NN-based green cognitive radio sensor networks for next-generation IoT

In any cognitive radio sensor networks (CRSNs), the secondary users (SUs) and primary users (PUs) share the opportunity to use the authorized frequency band. Here, the SU nodes can only transmit in the temporarily idle spectrum when it is not in use by any PU nodes. The proper estimation and detection of primary nodes are important for the energy-efficient spectrum access. The work basically presents a state of the art in implementing Bayesian-based convolutional neural network (CNN) for addressing the issue of energy constraints in next-generation Internet of things (Nx-IoT) using CRSN. Initially, we use blind source separation to extract the energy features and cyclic spectrum features of the signal and carry on the asymptotic autocorrelation calculation to the extracted signal. Finally, we construct the corresponding training set for CNN training and establish a suitable spectrum sensing model for Nx-IoT. Theoretical analysis and simulation results validate a suitable B-CNN spectrum sensing model along with energy-efficient cooperative communication between the SU nodes.