VANET distributed data fusion for traffic management

Abstract

In this article, we propose a distributed fusion algorithm to detect traffic congestion through the exchange of messages in vehicle network. This algorithm is based on the Dempster-Shafer theory that manages the uncertainties on data and sources of information. Each vehicle updates its database with local measurements (speed and interdistance) and information received from other vehicles and can calculate its route. Thanks to the collaboration, smart cars can avoid congested roads and take a better path to their destination. Several variants of the algorithm are studied and compared to a centralized approach through experiments carried out on the SUMO simulator using real urban road networks.

Collaborative Navigation Strategies Based on Vehicles Features in Connected Environments

“…Practically, in ideal condition (no blocked signals, no canyon effect), the range is up to 500 m in cities. In [23], authors have chosen to limit the range to 200 m when navigating in urban environments. In our method, as we exchanged information about obstructed lanes, that our city map is smaller than the one used in [23], and the vehicle's speed is limited to 30 km/h, we decided to limit the range of communication to 20 m, with the vicinity community membership defined in (4).…”

“…In [23], authors have chosen to limit the range to 200 m when navigating in urban environments. In our method, as we exchanged information about obstructed lanes, that our city map is smaller than the one used in [23], and the vehicle's speed is limited to 30 km/h, we decided to limit the range of communication to 20 m, with the vicinity community membership defined in (4). As the environment is crowded, and streets are small, when testing with large fleet of vehicles, this range is sufficient to access concerned vehicles when sending emergency messages.…”