Characterizing indoor wireless channels via ray tracing combined with stochastic modeling

We investigate the reliability of radio channel simulators in predicting channel responses throughout a well-specified environment. Indoor environments for which the geometric layout and material properties of surfaces are known lend themselves to such site-specific simulation. We assess the performance of this approach by comparing its predictions with measurements in a specific static environment. The good agreement on path loss, Ricean K-factor and RMS delay spread, over the set of paths measured and simulated, suggests that a well-designed radio simulator can be used reliably to predict system behavior. Typically, wireless channel models obtained through this or similar techniques do not capture the temporal variability in the channel response due to people movement in the environment. We treat the time-varying part of the channel response using stochastic processes. Using channel sounding experiments for several typical office scenarios, we show that autoregressive processes can be used to model the time-varying tap gains for several different motion scenarios.