Roadside units placement for total travel time minimization in I2X-enabled road networks based on a dynamic traffic flow model
TL;DR
This work proposes a heuristic approach for determining the placement of RSUs on a road network, while vehicle dynamics are modeled through a microscopic follow-the-leader interaction in which drivers continuously adapt their speed according to the distance from the vehicle ahead.
Abstract
This work addresses the problem of locating Road Side Units (RSUs) in vehicular networks with the aim of improving overall traffic efficiency. We propose a heuristic approach for determining the placement of RSUs on a road network, while vehicle dynamics are modeled through a microscopic follow-the-leader interaction in which drivers continuously adapt their speed according to the distance from the vehicle ahead. The RSUs share the collected information among themselves to cooperatively estimate the network state, which is then communicated to vehicles, allowing them to compute their fastest paths. We also assume that not all drivers comply with the provided recommendations, meaning that only a fraction of users follow the suggested routes. Since the aim is to improve network performance while minimizing the number of deployed units, the methodology adopts a nested optimization scheme: an outer module explores the optimal number k of RSUs, while an inner module identifies the best placement configuration for each k. To evaluate the effectiveness and the scalability of the proposed method, we tested two networks with different topologies and sizes. The results show a decrease in Total Travel Time (TTT) compared to the baseline scenario (i.e., without RSUs) of up to 36% on the smaller network and up to 20% on the larger one. Stochastic vehicle origin-destination pairs have also been considered.