FUZZY CONVOLUTION NEURAL NETWORK AND CONVERGENCE IMPROVED BAT OPTIMIZATION FOR AN ENERGY EFFICIENT AND SECURED SPECTRUM ACCESS IN COGNITIVE RADIO NETWORK

Abstract

Certain frequency bands are frequently used in traditional wireless communications systems but few are not utilized in specific time slots. The imbalanced spectrum allotments and service or traffic dynamics may be the reason for improper utilization of scarce spectrum resources. Cognitive radio (CR) technology is regarded as the prominent solution for mitigating spectrum under-utilization concern where secondary user can utilized the unused spectrum resources whenever primary users (PUs) is not using it. This is greatly achieved by the concept of spectrum sensing. A better-quality model work for optimum Spectrum Utilization of wireless systems with Cognitive Radio Networks has been suggested in existing work. Ensemble Adaptive Neuro Fuzzy Inference System (EANFIS) is greatly deployed for spectrum access control and Intrusion Detection is acquired through Enhanced support vector machine. The Particle Swarm Optimization (PSO) is also exploited for power allocation for secondary users. Conversely, there arises high ANFIS computational cost  because of complex structure and gradient learning. In addition, PSO may easily fall into local optimum in high-dimensional space and low convergence rate in iterative processes which are considered as PSO drawbacks. An enhanced model is presented for mitigating these problems through Fuzzy Filter Convolutional neural network (FFCNN) for spectrum access control and antecedents are regarded for spectrum selection of secondary user with trivial likelihood. Weighted fuzzy c means clustering plays a major role in detecting DOS and Replay attacks which is detected on basis of normal protocol operation behaviour, traffic flow and primary user access time. Kernel Cumulative Sum (KCUSUM) method is presented furthermore for detecting Primary User Emulation Attack (PUEA).  Convergence Improved Bat Optimization (CIBAT) is deployed too for secondary user’s power allocation and thereby diminishing energy consumption.  The suggested spectrum access control and power allocation model  efficiency are validated by experimental outcomes pertaining to throughput, packet delivery ratio and energy efficiency.