A learner based on neural network for cognitive radio

Abstract

Intelligence is a very important characteristic for cognitive radios (CR). Design of cognitive engine and application of artificial intelligence (AI) techniques are key to the implementation of this characteristic. Machine learning is one of the disciples in AI. This paper will propose a scheme of cognitive engine design, and use a learning algorithm based on neural network (NN) to implement a learner in the cognitive engine. A multilayer perceptron (MLP) neural network model will be introduced to ensure the convergence of the network, and problems on stop condition and overfitting will also be discussed. Finally, performance of the algorithm will be analyzed by simulations.