Comparative study of different V2V channel modeling strategies

Abstract

Abstract Vehicular Ad Hoc Networks (VANETs) enable Vehicle-to-Vehicle (V2V) and Vehicle-to-Everything (V2X) communications without central infrastructure. Wireless channel modeling is an important aspect of network performance evaluation and communication protocol optimization. This paper compares the received signal strength predictions of four deterministic propagation models, Free-Space, Two-Ray, Four-Ray, and Ray-Tracing, with real measurement data of the Berlin V2X dataset, and measures the error statistics and correlation coefficients. Our findings indicate that although the Free-Space model produced the highest correlation coefficient (0.79) and the lowest RMSE (4.34 dB), it missed out on signal variations. Ray-Tracing and Four-Ray models proved to be more successful in describing the signal dynamics and deep fading patterns, but with a little less statistical accuracy. Nevertheless, while the Ray-Tracing modeling is capable of capturing the dynamics of the measurements, the discrepancy in the locations of the “deep” fades indicates that Ray-Tracing modeling should be used with caution, both when studying a particular scenario or when using Ray-Tracing for creating datasets to train deep learning models.