Performance Analysis of Artificial Neural Network-Based Learning Schemes for Cognitive Radio Systems in LTE-UL

Abstract

Cognitive radio is widely accepted as a promising technology to intelligently manage the scarce radio resources and correspondingly select the optimal radio configurations. The process of cognition is challenging because of the trade-offs among response time, accuracy, available training samples, and NN structure complexity, which is a limiting factor for cognitive radio (CR) to achieve optimal configuration settings in real time. In this paper, a complex model of LTE uplink is analysed and a cognitive engine(CE) is introduced with ANN as an artificial intelligence technique. The CE is characterizing the achievable communication performance of all available secondary and primary users configurations. Furthermore, Suggesting the optimal radio configurations, taking into account the user requirements as well as the electromagnetic environment. Performance evaluation of the proposed ANN has revealed 60% improvement in accuracy and efficiency as compared to existing ANN models for the same parameters configurations.