Statistical Modeling of Intelligent Transportation Systems Communication Channels

Abstract

Intelligent Transportation Systems (ITS) is becoming an important paradigm, because of its ability to enhance safety and to mitigate congestion on road traffic scenarios. Realizing the fact that data collection scheme from in-situ test beds for large number of vehicles is always expensive and time consuming, before being employed in large scale, such safety critical system should be tested narrowing down the gap between real circumstances and analytical models in a simulation platform. It is evident that underlying radio wave propagation models can comprise the validity of large scale vehicular network simulation results. Vehicle to Vehicle (V2V) channels have higher dynamics due to rapidly varying topologies and environments which have significant impact on performance study of upper layer protocols and applications. In spite of the fact that few measurement based empirical channel models are present in the literature, they are not tested for large scale vehicular networks. Recently, IEEE has approved IEEE 802.11p which is the standard for Dedicated Short Range Communications (DSRC). During this MSc thesis, required amendment to the TNO’s in-house OPNET based 802.11p node model was made to meet the IEEE standard. To the best of author’s knowledge large scale simulation study of 802.11p nodes with realistic channel model is not present in the literature. Three different vehicular scenarios (i.e. suburban, highway, rural) were simulated with hundreds of IEEE 802.11p nodes in OPNET. For bringing down gap between real life behaviors of wireless channels in V2V communication a realistic Nakagami-m fading channel was developed to replace the standard OPNET propagation model. For the sake of implementing varying speed and separation distance, power spectrum and fading parameter-m were defined as function of velocity and separation distance respectively. Statistics were collected for large scale vehicular scenarios to evaluate performance of physical and higher layers. Retrieved scenario based results were compared with empirical channel models which showed fine resemblance between the mentioned models. Primarily, it was found that all the vehicles within the standard requirement for DSRC range of 1 kilometer may not receive packets. Finally, three different scenarios were collated which deduced that the multipath fading is more deciding factor for communication quality in contrast to the Doppler effect.