Autoregressive Modeling Approach for Non-Stationary Vehicular Channel Simulation

Abstract

A framework is proposed for long-term vehicular channel simulation based on the vector time-frequency autoregressive (VTFAR) model for a sparse parametric description of nonstationary multivariate random processes. Based on vehicle-to-infrastructure tunnel measurements, we estimate the VTFAR model parameters and validate the model by comparing the parametric and non-parametric spectra of the measured channel in terms of the delay spread and stationarity time. In addition, the VTFAR model stability is investigated and an approximation for the correlated scattering channel is proposed. The experimental validation shows a good agreement with RMSE of only 0.01 for the delay spread and 0.4 for the stationarity time. This approach provides an efficient alternative for non-stationary channel simulation that is measurement-based and computationally inexpensive.