Image Processing Reference
In-Depth Information
satisfiable results to the end user. Nevertheless, the design of a WSN system with its complex inter-
actions and system intricacies today requires in-depth knowledge and care for detail. As the field
of WSN continues to open up new application spaces, interaction and interdisciplinarity has to be
supported. Roemer et al. [] provide a broad overview of applications of these wireless networked
embedded systems to present a comprehensive design space of WSNs. The diversity in application
areas results in differing requirements for the system in terms of reliability, availability, quality of
sensor data, latency constraints, energy-efficiency, and longevity.
A framework has to be provided to determine detailed requirements of a platform, in order to
be able to aggressively optimize the system and provide users with a usable and satisfactory system
solution. he level of detail for accessing, programming, or optimizing system characteristics needs
to be different for a wide range of expertise of the system users.
WSN technology as many embedded systems has a long lifetime without the capabilities of easy
updates in the field. Thus, functional correctness at deployment time is of utmost importance.
However, numerous WSNs deployments fail to perform [,] or to work at all [], albeit ingenious
engineering efforts. A critical factor is that due to the novel environments and tight system integration
within, detailed models necessary for an understanding of the requirements are not available. As
an example in environmental monitoring, seasonal changes and its effect on plant growth heavily
influence sensed data, harvestable energy, and the properties of communication []. [].This results in
many system designs having to rely on either no, weak or simpliied assumptions.herealism of failing
and underperforming deployments has heavily influenced system design. Detailed provisions and
focuseddebug,maintenanceandmonitoringenhancementsimproveWSNreliabilityandperformance.
Researchers try to actively attack these problems by integrating sophisticated mechanisms into their
designs relying on in-depth knowledge [] and increased visibility by design []. Various tools attack
the debugging at the deployment site by listening to WSN traffic [], remote access to the sensor
nodes [], visual monitoring [], or integrating source-level debugging mechanisms for remote
debugging []. Nevertheless, WSN design today is tedious and requires attention to intricate details.
In order to deploy correct and performing systems, WSNs require tools for verification and val-
idation. Although validation in the form of testing can only determine the presence of errors and
not their absence, testing is the primary method employed in software, hardware, and system devel-
opment. Thus, we present testing tools and methodologies in the context of wireless networked
embedded systems, detail its peculiarities and explain according challenges.
11.1.1 Testing in the Context of Wireless Sensor Networks
WSNs integrate very different characteristics from various, previously unrelated fields: they are dis-
tributed systems built of embedded sensor nodes with wireless communication. Validation and
testing methodologies thus need to address the domain-specific intricacies and challenges.
11.1.1.1 Device Constraints
The sensor nodes have significantly benefited from Moore's law integrating minuscule micropro-
cessors, memory and radio chips at a reasonable price. Looking at the future there seem to be two
trends concerning computing and shrinking feature sizes: scale-up with small mote-class devices in
large numbers and scale-out with larger devices allowing for more functionality.
Larger devices allow for more ease-of-use in design and it well in tiered architectures. Not all appli-
cations require aggressive energy optimization and in select cases may work with powerful platforms,
especially when possibilities for energy harvesting [] are abundant. However cost and power con-
siderations in large-scale deployments favors mote-class devices, which require intricate attention in
design and testing.
Due to the tight resource constraints of embedded systems, software components and resource
usage have to be rigorously optimized. One of the most fundamental issues is the constraint of energy
