Channel modeling based on interference temperature in underlay cognitive wireless networks

Cognitive radio based dynamic spectrum access network is emerging as a technology to address spectrum scarcity. In this study, we assume that the channel is licensed to some primary (licensed) operator. We consider a sensor network with cognitive radio capability that acts as a secondary (unlicensed) network and uses the channel in underlay mode. The secondary network uses interference temperature model to ensure that the interference to the primary devices remain below a predefined threshold. We use hidden Markov model (HMM) to model the interference temperature dynamics of a primary channel. The HMM is trained using Baum-Welch procedure. The trained HMM is shown to be statistically stable. Secondary nodes use this trained HMM to predict the interference temperature of the channel in future time slots and computes the value of channel availability metric (CAM) for the channel. CAM is used by secondary nodes to select a primary channel for transmission. Results of application of such trained HMMs in channel selection in multi-channel wireless network are presented.