Support Vector Machine Based Spectrum Allocation Scheme for the Mobile Cognitive Radio Manhattan City Environments

Abstract

Cognitive radio (CR) is proposed as a critical means to reuse the primary spectrum in recent years. However, the cognitive node mobility has not fully researched for the mobile cognitive radio networks (CRNs). In this paper, a support vector machine (SVM) based spectrum assignment scheme is presented in the Manhattan city mobility environments, which takes the position and speed information of cognitive nodes into consideration during the spectrum availability prediction. Numerical results show good performance in the total spectrum utilization comparing with the traditional resource allocation algorithms.