Abstract:
Vehicular ad hoc networks (VANETs) have received considerable attentions
both from academics and industries. VANETs are promising for multi-hop
data delivery between a source and a destination vehicle (or a node) because of
their speci c characteristics such as high level of mobility in constrained and predictable
city/highway road networks. Multi-hop data delivery is useful for many
real-life applications, such as a driver or a passenger in the moving vehicle may
be interested to query for a sale in the shopping mall through xed locationbased
service provider, or to know about available parking spaces or current
tra c conditions of a region. In each cases, there is a need to handle real-time
tra c information to accurately transfer data between a source and a destination
node. Although, multi-hop data delivery is an well studied area, its existing
methodologies mainly focus on predicted/approximate tra c and cannot adopt
to dynamically changes of tra c condition. In this thesis, we address the problem
of handling real-time tra c information. To e ciently deliver the data
from a source to a destination we develop a novel methodology, Mobility Aware
Data Delivery (MADD) that considers global (predicted/approximate) and local
(real-time) tra c conditions. We develop three approaches: global window
based approach (MADD-G), local window based approach (MADD-L) and hybrid
approach (MADD-H). We carried out extensive experiment to demonstrate the
e ectiveness and e ciency of MADD-H and MADD-L with baseline approach
MADD-G and other techniques GPSR, GPCR and RBVT-R. Simulation results show that the MADD-H outperforms other approaches in terms of number of
hops, average delivery time and average delay.
