Statistical modeling framework of live spectrum observation data for opportunistic spectrum sharing

A statistical model of land mobile radio (LMR) voice traffic is developed from empirical RF Spectrum measurement data. This model builds upon previous work, and is used to generate synthetic voice traffic that closely follows the daily and weekly patterns of the measured traffic. The model is applied to a Dynamic Spectrum Access (DSA) simulation. A coexistence algorithm that takes advantage of the modeled channel statistics is presented that allows an opportunistic secondary user (SU) to share a channel with a primary user (PU). The algorithm shows a clear improvement compared to the basic Listen-before-talk scheme that has no knowledge of a PU's statistical traffic characteristics. Spectrum Opportunity Accessed and collision rate are used as metrics to compare the DSA coexistence techniques. We demonstrate the utility of a spectrum observatory system as being the integral part of this DSA framework, where the observatory continually monitors and models PU channel activity in order to provide the SU with useful statistical information about the PU's traffic characteristics.