Affiliations: Institut Telecom, Telecom SudParis, Evry, France | University of Valenciennes, Valenciennes, France | Inria Lille – Nord Europe, Lille, France
Note: [] Corresponding author: Thierry Delot, University of Valenciennes,
LAMIH UMR CNRS 8201 and Inria Lille – Nord Europe, Le Mont Houy, 59313
Valenciennes Cedex 9, France. Tel.: +33 3 2751 1956; Fax: +33 3 2751 1940;
E-mail: [email protected]
Abstract: This article focuses on data aggregation in vehicular ad hoc
networks (VANETs). In such networks, data produced by sensors or crowdsourcers
are exchanged between vehicles in order to warn or inform drivers when an event
occurs (e.g., an accident, a traffic congestion, a parking space released, a
vehicle with non-functioning brake lights, etc.). In the following, we propose
to generate spatio-temporal aggregates containing these data in order to keep a
summary of past events. We therefore use Flajolet-Martin sketches. Our goal is
then to exploit these aggregates to better assist the drivers. These aggregates
may indeed produce additional knowledge that may be useful when no event has
been recently transmitted by surrounding vehicles or when some knowledge about
the global demand may improve the decision that need to be taken at the vehicle
level. To prove the effectiveness of our approach, an extensive experimental
evaluation has been performed considering vehicles looking for an available
parking space, that proves the interest of our proposal. The experimentations
indeed show that the use of our aggregation structure significantly reduces the
time needed to actually find a parking space. It also increases the percentage
of vehicles finding such a resource in a bounded time in congested situations.
Keywords: Vehicular ad hoc networks, information sharing, spatio-temporal data aggregation, Flajolet-Martin sketches, communication protocols
