Impact of mobility design on network connectivity dynamics in urban environment

Abstract

Vehicular ad hoc networks have emerged as a promising research area for enabling various types of applications. However, VANETs are facing the biggest challenges in providing reliable communications due to unstable network connectivity caused by vehicle mobility. Several efforts have been carried out to analyze this problem. However, these efforts ignore the impact of intelligent mobility designs. This work highlights the relationship between intelligent mobility for transportation systems and connectivity dynamics. For this purpose, a framework of different mobility models is proposed based on the cellular automata (CA) approach to simulate vehicles mobility in a Manhattan two-dimensional network of roundabouts. To take into account intelligent aspects of mobility for drivers, a centralized path planning strategy based on the Bellman–Ford algorithm analyzes the travel time at road segments to provide the shortest paths for vehicles. Besides, we categorize three mobility models: In the first mobility model, vehicles are assigned the shortest paths in real-time with a periodic update. Each shortest path is defined as a set of turning movements (i.e., right, left, straight, …), where each turning step represents the driver’s decision at the next roundabout. At roundabouts, vehicles follow priority rules to avoid conflict with other traffics. The second mobility model is similar to the first one, but vehicles do have not the possibility to update their assigned shortest paths. The third model extends the first one by using traffic lights instead of priority rules at roundabouts. Extensive simulations based on both generated mobility traces and NS-2 analyze both network connectivity and reliability under several effective factors, including vehicles’ mobility model, vehicles’ speed, vehicles’ density, transmission range, and RSUs deployment strategy in terms of RSUs’ position and number. Several performance metrics of interest are introduced (such as packet delivery ratio, paths length, link duration, and end-to-end delay). The simulation results show that both the network performance and connectivity condition are sensitive to multiple factors, which reflect the V2V and V2I communications under varying conditions.