Research Activities
Digital Health
TBC.
Sports
Analytics
(In collaboration with A. Tommasi and C. Zavattari)I am actively involved in technological transfer activities, as co-founder of a spin-off of the University of Pisa that focuses on big data analysis for sports, a field of research also known as sports analytics. Sports Data Analysis is a hot topic nowadays, and in many disciplines it’s finding its way down from the pro players into a wider range of players. In particular, the spin-off focuses on tennis: its ultimate goal is to tune an innovative way of analysing games’s data, to provide the sport practitioner, both professional and amateur, a deep understanding of his/her playing style. The approach of the studied solutions are based on cutting-edge machine learning techniques to identify those aspects of the data that better characterize a player and his/her playing style. The resulting system learns how the player likes to play at his/her peak, and adapts itself to help him/her recreate those positive feelings, rather than the “count & report” approach normally adopted in this field by other available solutions.
Distributed Systems and Mobile Agents
Most of the research activity
developed in last few years focused on the study of mobile
and distributed systems.
Programmable
Matters. In collaboration with G. A. D Luna
(U. of Marseilles), P. Flocchini (Ottawa University), N.
Santoro (Carleton University) and G. Viglietta (ETHZ,
Zurich).
"Programmable matter" is typically viewed as a very
large number of very small (possibly Nano-level)
computational particles that are programmed to
collectively perform some global task by means of local
interactions. Such particles could have applications in a
variety of important situations: smart materials,
autonomous monitoring and repair, minimally invasive
surgery, etc. Such computational particles are under
development and there are multiple studies, mostly about
their implementation issues. So one can imagine to
tailor-make biological cells to operate as sensors and
actuators, as programmable delivery devices, and as
chemical factories for the assembly of Nano-scale
structures. This vision of building cheap microscopic
processing units is supported by the progress made in
manufacturing micro electronic mechanical components. Yet,
there is still no clear understanding which tasks can be
solved by micro fabricated computing elements (both
physically, and in the sense of computability) or which
problems can be solved efficiently. Our main goal is to
provide the algorithms to be used by those robots that we
expect that soon will be cheap and small enough, as a
result of the above mentioned efforts.
Control
and coordination of a set of autonomous mobile robots.
In collaboration with N. Santoro (Carleton University), P.
Flocchini (Ottawa University), P. Widmayer e M. Cieliebak
(ETH Zurigo), and V. Gervasi (Università di Pisa).
This research
topic focuses on the design and analysis of algorithms to
control and coordinate a set of autonomous mobile robots
that are allowed to freely move on a two dimensional
plane. The major goal of this work is to understand from a
computational point of view the relationship between the
power and the capabilities of the robots and the ability
of the team to accomplish the assigned tasks. One of the
outcomes of this work has been the design of a
computational model to describe a distributed system
populated by a set of mobile robots. The most innovative
feature that distinguished the model from the previous
ones present in the literature was the total asynchrony of
the interactions of the robots; in fact, in previous
models, the interactions between robots were modeled
mostly in a synchronous way. Other results concerned the
design of algorithms designed in the proposed model, that
allowed the team of the robots to solve tasks that are
common in robotics, such as pattern formation, gathering
or homing, flocking.
Black
hole search. In collaboration with N.
Santoro (Crleton University), P. Flocchini (Ottawa
University), and Stefan Dobrev (Ottawa University).
This research
topic relates to security in networked environments where
mobile agents compute. In these environments, in
fact, security is the most pressing concern, and possibly
the most difficult to address. The interest has been in
the presence of harmful hosts, called black holes: sites
that destroy any agent that might visit them without
leaving any observable trace. This kind of threat exists
not only on unregulated non-cooperative settings, such as
Internet, but also in environments with regulated access
and where agents cooperate towards common goals (e.g.,
sharing of resources or distribution of a computation on
the Grid). In fact, a local (hardware or software) failure
might render a host harmful. The focus of the research has
been to design efficient strategies for the agents to
locate and isolate these harmful presences, evidencing the
crucial importance of addressing security issues while
searching and exploring a network. Recently, the problem
of addressing a wider scenario --- the so called
"dangerous networks" --- is being tackled, where security
issues being dynamic and evolving over the time are
studied.
Safe-Routing.
In collaboration with N. Santoro (Carleton University), P.
Flocchini (Ottawa University), L. Pagli (Università di
Pisa), P. Widmayer (ETH Zurigo) e T. Zuwa (University of
Botswana, Gaborone).
This part of the
research activity has been devoted to the study of
fault-tolerant routing strategies in networks. In
particular, in systems using shortest-path routing
tables, a single link failure is enough to interrupt
the message transmission by disconnecting one or more
shortest-path spanning trees. The on-line recomputation of
an alternative path or of the entire new shortest
path trees, rebuilding the routing tables accordingly,
is usually rather expensive and causes long delays in
the message’s transmission. The focus has been on
the design of efficient distributed
algorithms to precompute, for each link in the tree,
a single non-tree link (the swap edge) able to
reconnect the network should the first fail. The
strategy, called point-of-failure swap rerouting is
simple: normal routing information will be used to route a
message to its destination. If, however, the next hop is
down, the message is first rerouted towards the swap
edge; once this is crossed, normal routing will resume.
The work focused on distributively identifying the swap
edges satisfying several optimazation criteria, such as
the distance between the point of failure and the
root, or the sum of the distances of all nodes below the
point of failure and the root.
Parallel Algorithms
This study is
focused on the design of parallel algorithms for the
Coarse-Grained Parallel Machine (CGM). In particular, we
study computation geometry algorithms and algorithms on
graphs. In collaboration with F. Dehne (Carleton
University), and A. Pietracaprina (Università di Padova).
You can find more information on my publications page....