In vehicular cognitive networks, a vehicle (secondary user) communicates with other vehicles or with the roadside units by borrowing licensed but unused frequency bands, a.k.a. white spaces, which are originally allocated to primary users. While doing so, secondary users make sure that the rights of the incumbent users are protected. In comparison to the cases with stationary secondary users, the probability that a primary user appears in the currently borrowed channel is much higher due to high mobility of vehicles. In order to avoid harmful interference to the primary users, a vehicle is required to vacate the currently borrowed channel and switch to another unused one. This vertical handover poses significant disruption to the ongoing communication due to the overhead induced by the channel handshaking. In this paper, we present a novel approach to reduce the number of vertical handovers. Specifically, we first introduce the White Space Vector (WSV) scheme which enables a simpler representation of the complex geographical white space information stored in spectrum usage databases. This new representation allows us to trim down the amount of data downloaded from the database required for adequately avoiding interference to the primary users. Furthermore, we present a new channel selection method leveraging the white space vector representation. We evaluate the number of vertical handovers from the point of view of accuracy and overhead, and show that WSV method can reduce the number of vertical handovers while also reducing the amount of data downloaded from the spectrum database without significant loss in accuracy.
White space vectors for channel selection in vehicular cognitive networks
2011-11-01
189464 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
CYCLOSTATIONARY FEATURE ANALYSIS OF CEN-DSRC FOR COGNITIVE VEHICULAR NETWORKS
British Library Conference Proceedings | 2013
|