Second Order Statistics and BER Performance Analysis of a non-WSSUS V2X Channel Model that Considers Velocity Variations

Abstract

Vehicular communication networks that may improve road safety are generating wide interest. As some vehicular networks use the radio frequency medium, channel estimation is crucial. Most channel estimation techniques are designed assuming the Wide Sense Stationary Uncorrelated Scattering (WSSUS) condition for the channel model. However, several measurement campaigns carried out around the world suggest that the WSSUS assumption might not hold in a vehicular environment. In this paper, we use simulations to compare the performance of two IEEE 802.11p-based Channel Estimation techniques in terms of Bit Error Rate (BER) with a channel model that is non-WSSUS. The selected channel model allows velocity variations on the mobiles. We compare the BER results of the estimation schemes varying the value of parameters related to the Doppler effect, namely, acceleration of the mobiles and the distance between them. Our results show that the nonstationarities caused by the Doppler effect are the main cause of the degradation of the system performance. We use the Four-Dimensional Time-Frequency Correlation Function to give prior insights on how the BER performance behaves. Also, we show that the BER performance of time-frequency estimation schemes, such as that of the Spectral Time Averaging technique, is strongly affected by the distance between the mobiles, but not by the accelerations of or the trajectory of these. On the other hand, we show that the BER performance of time interpolation only schemes, such as the Least Square technique, is not affected by the distance between the mobiles or the accelerations of them.