22-03-2013, 03:10 PM
Challenges in Wireless Sensor Networks
Challenges in Wireless Sensor.pdf (Size: 237.12 KB / Downloads: 264)
Abstract :
Software development for wireless sensor
networks requires novel programming paradigms and
technologies. This article describes the concept of a new
service oriented software architecture for mobile sensor
networks. With this architecture, a flexible, scalable
programming of applications based on an adaptive
middleware is possible. The middleware supports
mechanisms for cooperative data mining, selforganization,
networking, and energy optimization to
build higher-level service structures. The purpose of our
research activities is the development of a framework,
which radically simplifies the development of software for
sensor network applications.
Introduction
Deployment of a sensor network in a target area can be a
continuous process, for example to replace nodes with
depleted batteries or nodes that have been destroyed due to
environmental influences. In general, deployment establishes
an association of sensor nodes with objects, creatures, or
places in order to augment them with information-processing
capabilities. Deployment can be as diverse as establishing
one-to-one relationships by attaching sensor nodes to specific
items to be monitored [2], covering an area with locomotive
sensor nodes [7], or throwing nodes from an aircraft into an
area of interest [18]. Due to their large number, nodes have to
operate unattended after deployment. Once a sufficient
number of nodes has been deployed, the sensor network can
be used to fulfill its task. This task can be issued by an
external entity connected to the sensor network, such as a
user with a PDA, an aircraft flying by, or some device on the
Internet. Also conceivable are isolated, self-contained sensor
networks which are programmed to fulfill a certain sensing
task, whose result controls actuator nodes that are also part of
the network.
REQUIREMNTS OF SENSOR NETWORKS
Sensor networks have several requirements beyond those
needed by generic data networks. They must be able to
accommodate the Transducer Electronic Data Sheet (TEDS)
associated with each sensor. These networks often have more
stringent timing and synchronization requirements. Also, the
nodes must have lower power consumption and be smaller in
size than the products that support PC-to-PC networking.
Challenges in real time:
WSN deal with real world environments. In many cases,
sensor data must be delivered within time constraints so that
appropriate observations can be made or actions taken. Very
few results exist to date regarding meeting real-time
requirements in WSN. Most protocols either ignore real-time
or simply attempt to process as fast as possible and hope that
this speed is sufficient to meet deadlines. Some initial results
exist for real-time routing. For example, the RAP protocol [1]
proposes a new policy called velocity monotonic scheduling.
Here a packet has a deadline and a distance to travel. Using
these parameters a packet’s average velocity requirement is
computed and at each hop packets are scheduled for
transmission based on the highest velocity requirement of
any packets at this node. While this protocol addresses
realtime, no guarantees are given. Another routing protocol
that addresses real-time are called SPEED [2]. This protocol
uses feedback control to guarantee that each node maintains
an average delay for packets transiting a node. Given this
delay and the distance to travel (in hops), it can be
determined if a packet meets its deadline (in steady state).
However, transient behavior, message losses, congestion,
noise and other problems cause these guarantees to be
limited. To date, the limited results that have appeared for
WSN regarding real-time issues has been in routing. Many
other functions must also meet real-time constraints
including: data fusion, data transmission, target and event
detection and classification, query processing, and security.
New results are needed to guarantee soft realtime
requirements and that deal with the realities of WSN such as
lost messages, noise and congestion. Using feedback control
to address both steady state and transient behavior seems to
hold promise. Dealing with real-time usually identifies the
need for differentiated services, e.g., routing solutions need
to support different classes of traffic; guarantees for the
important traffic and less support for unimportant traffic. It is
important not only to develop real-time protocols for WSN,
but associated analysis techniques must also be developed
(see the section below on Analysis).
Programming Abstractions:
A key to the growth of WSN is raising the level of
abstraction for programmers. Currently, programmers deal
with too many low levels details regarding sensing and node
to node communication. For example, they typically deal
with sensing data, fusing data and moving data. They deal
with particular node to node communication and details. If
we raise the level of abstraction to consider aggregate
behavior, application functionality and direct support for
scaling issues then productivity increases. Current research in
programming abstractions for WSN can be categorized into 7
areas: environmental, middleware APIs, database centric,
event based, virtual machines, scripts and component-based.
As an example, consider an environmental based abstraction
called EnviroTrack [3].