One of the challenges that the wireless industry faces today is to provide mass multimedia
services to mobile and portable devices at low cost, making them affordable to users and
profitable to operators. The most representative service is mobile TV, which is expected to
become a key application in future wireless networks. Nowadays cellular networks cannot
support large scale consumption of such services, and newly deployed mobile broadcasting
networks are very expensive to deploy due to the large investment in infrastructure required
to provide acceptable coverage levels.
This dissertation addresses the problem of providing a cost efficient provisioning of
mass mobile multimedia services with existing wireless infrastructure for digital broadcasting
and cellular systems. State-of-the-art mobile broadcasting (DVB-H) and cellular
(evolved 3G+ networks with HSDPA and MBMS) commercial technologies have been
considered. Our approach is DVB-H centric. The main paradigm proposed to provide
lower cost services is to avoid deployment a DVB-H network with high capacity and
full area coverage from the beginning. Instead, it is proposed to perform an incremental
DVB-H network deployment that follows the user demand. Within this context, the
cellular network is essential to avoid over-dimensioning the broadcast capacity, and also
in low populated areas until the deployment of a DVB-H transmitter or gap-filler becomes
convenient.
The main technical solution proposed in this dissertation is to perform a multi-burst
protection of the transmission in DVB-H using Raptor coding. The idea behind is to exploit
the time diversity of the mobile channel to improve the robustness of the transmission, and
thus the coverage level, at the expense of an increased network latency. Contrarily to
most technical enhancements proposed in the literature, the introduction of this technique
would be backwards-compatible with existing networks and terminals, since it could be
introduced as a software update either at the link or the application layer.
On the other hand, this dissertation also investigates the potential efficiency improvement
that can be achieved in a hybrid DVB-H/3G+ system when the cellular network is
used to repair errors of the broadcast DVB-H transmissions. Raptor coding is also considered
in this case, but at the application layer; since it enables an easy and efficient
implementation of the repair mechanisms in such hybrid systems.
In the dissertation we evaluate the potential gain that can be obtained with multi-burst
forward error correction (FEC) in DVB-H compared to the conventional intra-burst FEC
mechanism by means of simulations, lab and field measurements, and theoretical studies.
Our results show that it is feasible to significantly increase the area coverage and the service
quality in already deployed networks, as well as to reduce the investment in infrastructure
required to deploy new networks. The gain is especially important for services where a
large latency is not an issue, such as file download services. In these cases the area coverage
can easily double, and even treble under certain conditions. The gain increases for large
files, robust coding rates and for vehicular users due to their high mobility. Furthermore, for
this kind of services without stringent time constraints it is possible to reduce the delivery
cost by half combining a DVB-H network with a 3G+ cellular network.
For DVB-H streaming services the gain due to multi-burst FEC is limited if low zapping
times are desired. In this case it is necessary to combine the multi-burst coding with
techniques aimed to reduce the zapping time from the perception of the users. However,
it is possible to significantly increase the robustness of the transmission transmitting few
services heavily protected. In particular it is possible to double the area coverage for vehicular
users and up to 50% for pedestrian users. The proposed technique would also increase
the service quality in already covered areas, removing practically all residual errors of the
transmission.