A radio environment map construction scheme with hidden Markov Model based spectrum occupancy prediction

Abstract

A radio environment map (REM) construction scheme with hidden Markov model (HMM) based spectrum occupancy prediction is presented in this paper. The predicted spectrum occupancy state can be shown visually in the REM, which is used to find the opportunity to access a certain channel. Firstly, the HMM is studied to predict the spectrum occupancy state at the future time in a unsupervised manner. Secondly, according to the predicted result from the HMM, we construct the REM by the extrapolation method based on wireless propagation loss models to reflect the potential spectrum occupancy status and make a decision on accessing the channel or not. Simulation results show that the proposed scheme has a good prediction performance and the constructed REM is helpful for the secondary user to access the desired channel.