Spectrum Occupancy Modeling and Prediction for Efand#64257;cient Dynamic Spectrum Access in Cognitive Radio Networks

Abstract

Duetoever-increasingtrafand#64257;candnon-availabilityofnewspectrum,efand#64257;cientuseofexistingspectrum has become a necessity. Under the current system of static spectrum access, the spectrum allocationutilization ratio is very low. Using dynamic spectrum access (DSA), cognitive radio (CR) systems promise opportunistic use of unused spectrum space in the licensed band to signiand#64257;cantly increase the spectrum efand#64257;ciency. Investigation of spectrum occupancy through spectrum occupancy measurements aswellasspectrumsensinginbothlicensedaswellasunlicensedspectrumbands,providethenecessary information of spectrum usage that is very useful in the CR deployment. However, detailed measurements and analyses intended to quantify the performance of one particular system usually cannot be applied directly to other cases. So in this context, it is required for CR systems to stochastically model the spectral activity of the licensed system through accurate spectrum occupancy characterization. The existingmodelsarehighlydependentonthespatio-temporalvariationsoftheprimaryuser(PU)activity asthestatisticalbehaviorofthePUchangeswithrespecttothelocation,spectrumband,andthevarying load time. In this context, a generalized spectrum occupancy model with tunable parameters is required which can be and#64257;t into any spectrum data irrespective of the varying aforementioned conditions. Through spectrum occupancy modeling, sometimes it becomes difand#64257;cult to ascertain a-priori the pattern of the spectrum usage of the PU due to its stochastic behaviour. Hence, it becomes necessary to estimatethefuturespectrumavailabilityofaPUchannelthroughsomeforecasting/predictivetechnique for enhanced quality of experience of secondary users (SUs) and efand#64257;cient DSA. newline