|
Wireless and Mobile Computing 1. Using (mobile) Agents in Mobile Computing and Sensor
Networks Recently,
a new approach for learning features of a domain of interest has been
proposed. The main characteristic of the novel approach is the fact that
global information about the domain is obtained by combining in some clever
way local information gathered by independent agents. In fact, for a large
number of practically-relevant domains the following paradigm can be used:
(1) a large number of agents, each with a specific mandate is being sent into
the domain, (2) each agent learns a given characteristic or feature of the
domain, (3) a subset of the agents is recovered and debriefed. Somewhat
surprisingly, for many domains it is possibly to recover a strict subset of
the agents and still obtain “full” knowledge about the domain. It
would be very useful to implement strategies for 1-3 above for a number of
particular domains arising in various practical applications. Of a particular
interest is the area of mobile computing and wireless networks. 2.
Mobility Issues and Location Management in Wireless Networks
In
mobile computing systems the connectivity of the underlying network does not
change in the absence of communication link or host failure. On the other
hand, mobile hosts are capable of moving between different locations while
maintaining their connectivity to the network - for example, via cellular
connection or a packet radio network. In standard terminology, a mobile host
is an entity capable of both communication and of performing
“local” computation. This is the main characteristic that
differentiates a mobile host from communication-only devices such as pagers
or portable dumb terminals. Mobility of hosts introduces a new set of issues
that were not present in networks and distributed systems with static hosts
only. For example, to deliver a message to a mobile host, it is necessary
that the current position of the destination host be first identified within
the network. As the hosts move, the physical connectivity of the network
changes. Hence, any logical structure that many network algorithms exploit -
for example, spanning trees, path covers, etc., cannot be statically mapped
to a set of physical connection within the mobile network. Mobile hosts have
severe resource constraints in terms of limited battery life and, as a
consequence, often operate in “sleep” mode, with sporadic bursts
of network activity, or entirely disconnected from the network. The
communication between a mobile host and the remainder of the network occurs
via a wireless medium. Such a medium physically supports broadcast communications
within a specified region, commonly referred to as cell. These aspects are
characteristic of mobile computing and have to be considered in the design of
algorithms, centralized or distributed, for these networks. 3.
Cost- and Power-Aware Localized Routing in Wireless Networks
Wireless
devices must operate for a long period of time, relying only on their battery
power. While many developers have looked at extending the life of a mobile
system from a hardware point of view, such as directional antennas and
improving battery life, power based routing is a relatively new concept in
wireless networking. Until recently most routing protocols in wireless
networks have concentrated mainly on establishing routes, and maintaining
these routes under frequent and unpredictable changes in network topology.
The concept of using routing to minimise power usage has only recently been
looked at and it has shown to be moderately successful. It has been proposed
that routing packets in a power aware method will complement hardware based
methods of extending the network’s life. The metrics that have so far
been devised to minimise power can be grouped into two main groups,
power-aware and cost-aware metrics. Power-aware metrics aim to minimise the
total power needed to route a message between two different locations while
cost-aware metrics look at the methods which extend the node’s
battery’s life time. The aim of the project is to further develop
power-aware and cost-aware metrics, which have been devised to minimise power
loss and maximise battery life of wireless units. Once suitable metrics have
been developed it is intended that the power-aware and cost-aware methods
will be combined to produce a power and cost aware routing method. 4. Self-Organising Protocols for Wireless Sensor Networks In
a wireless sensor network, energy conservation is the primary design goal.
Research shows that in a low-energy radio network, the energy consumed by
receiving and listening (attempting to receive) messages is
of the same order of magnitude as transmitting them. The most efficient way
to save energy is keeping sensor nodes turned off as long as possible. These
sleeping-or-awaking nodes need the capabilities of self-organisation and
re-organisation to adapt to dynamic environment and network settings. This
work address issues of energy efficient self-organisation in sensor networks.
The work also deals with situations in which the network needs to efficiently
adapt in catastrophe scenarios by maintaining reasonable energy levels that
keep the network active for the longest period of time. 5. A Biomimetic
View of Large-scale Sensor Network Systems
The
massive initial deployment of sensor nodes provides the basics necessary for
the organization of a primitive community.
The sensors then mimic births and deaths in a biological system as
they are activated (born) in waves, or generations, each generation adding
its learned collective experience to the genetic base and passing the
enhanced genetic information along to the next generation. The new genetic material will be local in
scope allowing regional properties in the evolutionary process. Thus, sensors
belonging to the new generation and situated in different locales of the
community will be “born” with differing genetic materials. This will allow the sensor net to perform
efficiently in a non-uniform environment. The regimen of updating the genetic
material from one generation to the next ensures that, just as a biological
system, the sensor network evolves and, in the process, may change its
strategy, priorities, and methods. Evolution is game-theoretic and is guided
by the maximization of a global objective function, based on local data only.
Learning is key to ensuring that individual sensors,
or groups of sensors specialize in a way that furthers the interests of the
sensor network community. At one level of abstraction, collective
intelligence can be defined as the ability of the sensor network community,
given a specific current network state, to effect a transition to a next
state that better serves the overall community goal. The localized nature of
communication in sensor networks dictates a decentralized approach towards
the development (and evolution) of collective intelligence. In a
decentralized approach, neighboring sensors evolve
a local collective intelligence. The composition of local collective
intelligence is what gives rise to a communal collective intelligence. One
fundamental problem is how to guarantee the correctness of communal
collective intelligence, up to community goals, given the relatively
autonomous nature of local collective intelligence evolution? To this end,
one scenario is that an arbitrary group of neighboring
sensors would initiate a locale, and negotiate, as a single entity and as
warranted, local goals for their locale for x amount of time. These local
goals would be, of course, a function of the overall community goal. During
the negotiated time, the sensors effect state transitions that
‘better’ serve their negotiated locale goal(s), triggered by
input stimuli from both inside and outside the locale. 6. Distributed Coalition Planning and
Decision Making
Our
research in this particular area proceeds along two distinct directions.
First, we are interested in one specific aspect of federating resources
namely the establishment of coalitions both in a game-theoretic sense and
from a generational-learning and service-centric perspective. We view
coalitions as being avenues for maximizing a given (often global) objective
function subject to (mostly local) constraints. Our work focuses mostly on
wireless sensor networks and coalitions of networks subject to functional
mobility as opposed to physical mobility. Indeed, in sensor networks
populated by fixed sensors, one can define strategies for functional
migration very much akin to physical mobility, except that it is not visible
to the adversary. In a military environment this translates into low
detection probability. 7. Federating Autonomous Sensor Networks An important component of our research is
motivated by the need to use the inherent capacity of sensor networks for
data collection, surveillance and target tracking as a key ingredient for
establishing ubiquitous monitoring and control capabilities in support of
civilian and defence applications. Indeed, a single sensor network cannot
satisfy the broad spectrum of application requirements, especially when these
requirements change drastically along the dimensions of time, space, and
context. On the other hand, deploying numerous sensor networks in an area of
interest may be infeasible. The goal is to develop a new sensor network
system that will act as a distributed service provider. To build such a
distributed system, we are looking at innovative sensor network system
architectures that will facilitate rapid self-organization and dynamic
reconfiguration of component sensor networks in support of adaptive service
deployment, composition, and federation to cover the dynamic needs of
numerous applications. 8. Protecting with Sensor
Networks Sensors, in one form or another
have always been a component in physical security systems. Usually such sensors
are configured on a perimeter, or perhaps on concentric perimeters. This project
is motivated by the realization of the fact that the probability of detecting
an intruder is a Quality of Service (QoS)
parameter. This implies an interesting tradeoff
between the amount of resources that a defender can muster and the QoS (in terms of probability of detection) that they get.
We have studied this trade-off for a family of structures with an axial
design reminiscent of a snowflake. We show that such a structure presents
interesting qualities. The relations geometrically deduced in the paper
provide a form of sensitivity analysis. 9. National
Security Currently in debate is a suggestion that we use military
personnel to patrol borders. We view intelligent sensor network communities
as an ideal basic technology for establishing an intelligent
border-monitoring infrastructure. Specifically, intelligent sensor network
communities can be used to implement a programmable
sensory infrastructure. Diverse surveillance,
command, and control applications can
then be supported on top of this sensory infrastructure. The general-purpose
nature of the sensory infrastructure would make this a highly economical
solution. Additionally, the principles of detection and estimation theory
suggest that high detection accuracy invariably demands a large number of
sensory data points; a large-scale densely populated sensor network community
is an ideal solution for economically and reliably collecting (and
intelligently filtering) large quantities of sensory data points. Conversely,
the scarce human intelligence resource is best utilized at the high-level
application layer to drive analysis and decision-making.
|
