Summary:
In this paper, we are, to the best of our knowledge, the first to question the
robustness of parking assistance systems to non-cooperative drivers’ behaviors,
which deviate from the purely altruistic paradigm of always truthfully sharing
the cached information with encountered vehicles. Hence, we let nodes misbehave
and study how this affects fundamental performance indices such as the parking
search time and the distance of the acquired parking spots from the drivers’
travel destinations. The dual question from a driver’s viewpoint is whether
nodes do have incentives to misbehave in that misbehaving lets them achieve
better search times and/or parking spot-destination distances. Two intuitive
instances of misbehaviors are considered. In the first one, nodes defer from
sharing parking information with other vehicles essentially acting as free
riders. In the second one, they deliberately falsify information about the
parking spots’ status (selfish liars), i.e., spots close to a misbehaving
vehicle’s destination are advertised as occupied whereas all others as vacant.
The two misbehaviors essentially impair in different manner the amount and
accuracy of information that is disseminated across the network. Finally, we
take into consideration the case where misbehaving nodes in a centrally
assisted parking search system try to bypass sometimes system’s procedure, when
destinations are uniformly distributed, in order to obtain a better spot than
the one the assigned to them.
Keywords:
Parking assistance systems, Non-cooperative opportunistic dissemination, VANETs, Full memory use, Random memory use
