14 junho 2019, 15:52
•
António Manuel Raminhos Cordeiro Grilo
Although I have not officially proposed any master theses, in case there are interested students, the following subjects are available (I have more topics, but there are the highest priority ones):
- Optimization of Mesh Networks of Flying Base stations Using Neural Networks:
Our group at INESC has been working on optimization of ad-hoc networks, where ground nodes connect to the cloud through UAV base stations. The objective is to find optimal positions for the flying base stations, based on the propagation characteristics of the environment, the location and traffic requirements of the ground nodes. This is useful in scenarios of natural disaster, military applications, as well as when the capacity of the cellular network is not enough to cope with some temporary crowded event. Until now, the optimization algorithms were based on classical paradigms, such as Genetic Algorithms. To perform optimization based on Neural Networks is an emerging research area.
- Fog Computing Optimization in 5G Mobile Networks:
Fifth generation (5G) cellular network promises to
offer to its users transmission speeds in the order of the Gbit/s and submillisecond
latency, enabling resource heavy applications, such as Mobile Gaming,
intelligent Distributed Camera Networks and Internet of Things. However, the
current cloud based computation and data delivery model do not allow the
required quality of service (QoS) guarantees to be efficiently harnessed, due
to the number of hops of wired networks between the 5G base stations and the
cloud, that leads to a significant increase in latency. Forwarding all the data
generated by devices directly to the cloud may devour the bandwidth and lead to
congestion. Therefore, it is necessary that processing be hosted near the
devices, close to the source of the data, so that the high speed transmission
of 5G can be utilized and data can be processed and filtered out by the time it
reaches the cloud. This is called Fog Computing. This thesis will focus on the optimization of Fog resource management, building upon previous work and simulation software developed at INESC.
- On-Site Propagation Model Parameterization based on
Fuzzy Bayesian Nets:
Topology optimization of networks of UAV base stations must
rely on propagation models, which try to approximate the characteristcs of the
environment. A priory parameterization of propagation models based on general
known characteristics of the environment, usually lead to huge errors, since
reality never follows the model exactly. If the propagation characteristics of
the environment can be learnt fast enough while UAVs are operating, this will
contribute to the improvement of the accuracy of propagation estimates, and
hence to more optimized topology solutions. In this project, Fuzzy Bayesian
Nets will be used for on-site parameterization of the propagation models.
Testing of the algorithms will be based on simulation scenarios, and also real
measurements on the field using WiFi.
Regards,
A. Grilo
