abstract
-
Smart scheduling can be used to reduce infrastructure-to-vehicle
energy costs in delay tolerant vehicular
networks (Hammad et al., 2010).. In this thesis we show that by combining
this with vehicle-to-vehicle (V2V) forwarding, energy efficiency can
be increased beyond that possible in the single hop case. This is
accomplished by having the roadside infrastructure forward packets
through vehicles which are in energy favourable locations. We first
derive offline bounds on the downlink energy usage for a given input
sample function when V2V forwarding is used. Separate bounds are given
for the off-channel and in-channel forwarding cases. These bounds are
used for comparisons with a variety of proposed online scheduling
algorithms. The paper then introduces online algorithms for both
fixed bit rate and variable bit rate air interface options. The first
algorithm is based on a greedy local optimization (GLOA). A version of
this algorithm which uses a minimum cost flow graph scheduler is also
introduced. A more sophisticated algorithm is then proposed which is
based on a finite window group optimization (FWGO). Versions of these
algorithms are also proposed which use in-channel vehicle-to-vehicle
scheduling. The proposed algorithms are also adapted to the variable
bit rate air interface case. Results from a variety of experiments
show that the proposed scheduling algorithms can significantly improve
the downlink energy requirements of the roadside unit compared to the
case where vehicle-to-vehicle packet forwarding is not used. The
performance improvements are especially strong under heavy loading
conditions and when the variation in vehicle communication
requirements or vehicle speed is high.